available for freelance & full-time
Full Stack Developer with 2.5+ years building scalable products on React, Next.js, Node.js and NestJS — plus Generative AI: LangChain.js, RAG pipelines, vector databases and MCP servers for AI-driven products.
2.5+
years shipping production code
10+
projects across web & mobile
3
platforms — web · mobile · desktop
RAG
systems, MCP servers & agents
[ 01 ]about
education
Master of Computer Applications (MCA)
Varanasi, India
Bachelor of Computer Applications (BCA)
Varanasi, India
Computer Operator & Programming Assistant
1-year ITI
I build products the way I like my APIs — fast, typed and predictable. From a quiz-based learning platform serving web and mobile, to a production RAG system answering questions over internal documents.
Over the last 2.5+ years I've worked across the whole surface of a product: designing secure REST APIs with authentication and optimised queries in NestJS and Express, engineering front-ends with Next.js, Tailwind and Material UI, and modelling data in PostgreSQL, MongoDB and MySQL.
The newest layer of my work is AI engineering — LangChain.js agents, retrieval-augmented generation, embeddings and MCP servers that let models safely use real tools. That combination — solid full-stack craft plus applied LLM systems — is what I bring to every build.
Based in
Noida, India
Experience
2.5+ years, full-time
Currently
Senior Full Stack @ Digixito Media
Focus
MERN + Generative AI products
beyond the code
[ 02 ]skills
A working toolbox, not a keyword dump — everything below has shipped to production in the last 2.5 years.
6 domains · 30+ tools
/01
Interfaces that feel instant
/02
Secure, well-shaped APIs
/03
Modelled for the query, not the ORM
/04
LLMs that use real tools
/05
Shipped, monitored, deployed
/06
Daily drivers
[ 03 ]experience
Digixito Media Pvt. Ltd. · Noida, India
Dec 2024 — Apr 2026
Hindsol Software Pvt. Ltd. · Varanasi, India
Dec 2023 — Nov 2024
[ 04 ]selected work
Live from the database — this grid is managed through the dashboard, so new work shows up here the moment it's added.
// nothing here yet
Add projects from the dashboard and they appear instantly.
[ 05 ]ai lab
The second half of my craft: making LLMs useful in production — grounded in your data, wired to your tools.
rag_pipeline.flow — how an answer gets grounded
query
user intent
embed
vectorise
retrieve
vector store
reason
LLM + tools
answer
with citations
Ingestion, chunking, embeddings and retrieval tuned per corpus — the pattern behind ShipGPT.
Model Context Protocol servers that expose real tools and data to LLMs safely and predictably.
LangChain.js agents that plan, call tools, self-correct and hand off — not just chat.
Prompt scaffolds, structured outputs, streaming responses and evaluation loops in production apps.
case in point — ShipGPT
A production RAG system that answers questions over internal documents with cited context — chunking strategy, vector retrieval and prompt scaffolds built to survive real users.
[ 06 ]services
Web, mobile, desktop or AI — one engineer across the whole surface. Hover a row to see what's inside.
01
Fast, SEO-healthy web apps and sites
Marketing sites, dashboards and full SaaS builds — server-rendered Next.js front-ends with NestJS or Express backends, auth, payments and admin panels included.
02
Cross-platform mobile apps in React Native
One codebase, both stores. Shared APIs and auth with the web build — like mytreks.ai, which runs as both a website and a mobile app.
03
Desktop tools that talk to the cloud
Cross-platform desktop applications with web technology — local-first flows, installers and cloud sync where it matters.
04
LLM features that survive production
RAG pipelines over your documents, MCP servers, agentic workflows and chat copilots — with embeddings, evaluation and streaming UX done properly.
05
The part users never see — until it breaks
Secure REST APIs with JWT auth, clean schema design and optimised queries across PostgreSQL, MongoDB and MySQL. Documented, tested, deployable on AWS, Vercel or Render.
[ 07 ]contact
App, web platform, desktop tool or an AI feature that needs to be real — send the shape of the problem and I'll reply within 24 hours.
tip: messages land straight in the dashboard inbox — this site runs its own Next.js + PostgreSQL backend.
new_message.form
POST /api/messages